Gérard,

Excellent, glad to have been able to help,

George

baaann wrote:I can confirm GeorgeB's results(including fonts)

Noted!

baaann wrote: although I classify tests 10 + 11 as fail due to the text not being repeated on each page.

baaann,
If it is repeated, then it fulfills the HTML 4 expectations but the spec is loose and the conditions are relaxed because of a "may" word.

{
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
}
http://www.w3.org/TR/html4/conform.html

and RFC2119 states that "may" means optional.

baaann wrote:I did the tests using the PDF printer, Konqueror 4.6.5, KHTML and Gentoo.

Noted!

salutations amicales de Montréal,
Gérard Talbot

GeorgeB wrote:

GTalbot wrote:

GeorgeB wrote:10. Prints but appears to do so incorrectly (I can't make sense of what the expected result is but konqeror is different from ffx)
11. Ditto

When you have a long table to print and if you declare/create a table footer, then it should be printed at the bottom of each page for such long table.

Right, thats what I thought but the wording on the test page itself seemed to allow it to either be printed or not as long as its consitent, in that case its a fail.

It should be printed but it (<thead> text data or <tfoot> text data) may or may not be repeated onto the other pages. I am now convinced that the test pass conditions are not ideal.

"
That the test instructions are accurate, precise, simple, and self-explanatory. Your mother/husband/roommate/brother/bus driver should be able to say whether the test passed or failed within a few seconds, and not need to spend several minutes thinking or asking questions.
"
CSS 2.1 test suite, Review Checklist

GeorgeB wrote:

GTalbot wrote:

GeorgeB wrote:12. + 13. Looks OK to me - I don't understant cyrillic but do have fonts for it for some reason.

I wonder which font you use, have set up for cyrillic. When I right-click on "DejaVu Serif" font name in Font installer, then choose "View in Font Reader" and then choose "Unicode Block: Cyrillic", I see all of the cyrillic characters that "DejaVu Serif, classic" font can display.
I will probably file a bug on this as I still do not understand why the test, when tried in Konqueror 4.6.5/KHTML, produces so many pairs of undisplayed/undisplayable characters.

I'm using DejaVu Sans as the default font - looking at the font in Font Reader it appears to have more cyrillic characters. Droid Sans also has the right glyphs for the test.

All the best,
George

DejaVu Sans and Droid Sans: noted! I'll try these later today.

salutations amicales de Montréal,

Gérard Talbot

GeorgeB wrote:12. + 13. Looks OK to me - I don't understant cyrillic but do have fonts for it for some reason.
(...)
I'm using DejaVu Sans as the default font - looking at the font in Font Reader it appears to have more cyrillic characters. Droid Sans also has the right glyphs for the test.

George, You were right. DejaVu Sans is better for such 2 testcases.
I do not have Droid Sans; I'll try to install it later.

-------------

So, I have updated the results of those 13 testcases both in the detailed implementation report and in the basic implementation report.

I have linked the results of those 13 testcases to this thread.

E.g.:
http://www.gtalbot.org/BrowserBugsSection/Konqueror4Bugs/Konq-IR-CSS21TestSuite-79.html has 8 of your testcase results.

I will submit the Implementation Report probably tomorrow.

For now, here are the final results I end up with for Release Candidate 6 (March 23rd 2011) for Konqueror 4.6.5 with KHTML rendering engine:

7981 testcases are PASSED

1345 testcases are FAILED

46 testcases are INVALID (more than 46 if we apply strict rules)

4 testcases are UNCERTAIN

When we exclude invalid and incertain testcases, then:

7981 / (7981 + 1345 == 9326) ==

85.58%



The results of Chrome 6, Opera 10.70, Safari 5.0.2 and Firefox 4.0beta6 for RC1 were:

Chrome 6
--------
# Chrome 6.0.472.63 Windows 7
# Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US) AppleWebKit/534.3 (KHTML, like Gecko) Chrome/6.0.472.63 Safari/534.3
# http://test.csswg.org/suites/css2.1/20100917/
which was RC1.
8301 passed
1217 failed
and many were skipped due to various issues.
8301 / (8301 + 1217) = 87.21%

Opera 10.70
-----------
# Opera 10.70 (build 9067) Windows 2000
# Opera/9.80 (Windows NT 5.0; U; en) Presto/2.6.37 Version/10.70
# http://test.csswg.org/suites/css2.1/20101001/
which was RC2, I believe.
8704 passed
961 failed
68 invalid
============
9733

8704 / (8704 + 961) = 90.05%


Safari 5.0.2
------------
# Safari 5.0.2 MacIntel
# Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_6_4; en-us) AppleWebKit/533.18.1 (KHTML, like Gecko) Version/5.0.2 Safari/533.18.5
# http://test.csswg.org/suites/css2.1/20101001/
which was RC2, I believe.
16592 passed
2741 failed
89 ? (unknown)
41 skipped
3 skipped (invalid)
==================
19466

16592 / (16592 + 2741) = 85.82%


Firefox 4.0 beta 6
-------------------
# Firefox 4.0b6 Ubuntu 10.04
# Mozilla/5.0 (X11; Linux x86_64; rv:2.0b6) Gecko/20100101 Firefox/4.0b6
# http://test.csswg.org/suites/css2.1/20100917/
which is RC1.
18188 passed
691 failed
494 invalid
84 unknown
=============
19457

18188 / (18188 + 691) = 96.34%

Internet Explorer 9
-------------------
# Internet Explorer 9.0 on Windows 7
# Internet Explorer 9.0 (Trident 5.0) on Windows NT 6.1
# http://test.csswg.org/suites/css2.1/20110323/
#IE9 version 9.0.8112.16421 (RTM)

Pass: 18528 - 98.78%
Fail: 228 - 1.21%
Invalid: 2
N/A: 4

regards, Gérard

GTalbot wrote:I will post the results for each of those 13 testcases and link it to this thread.

GeorgeB,

I have posted the results of each of those 13 testcases and have linked those to this thread.

E.g.:
This page
http://www.gtalbot.org/BrowserBugsSection/Konqueror4Bugs/Konq-IR-CSS21TestSuite-79.html
has 8 of your results.

salutations amicales de Montréal,
Gérard Talbot

Ah thanks for that, those must have taken a lot of work to put together.

GeorgeB wrote:Ah thanks for that, those must have taken a lot of work to put together.

8 months!

And I filed 33 bug reports on the failures

salutations amicales de Montréal,
Gérard Talbot

Thanks for that list of bug reports, can you remember off the top of your head if any of them are related to my reports? If so I'll subscribe to those ones. If not I'll try to get round to reporting them myself.

GeorgeB wrote:(...) any of them are related to my reports? If so I'll subscribe to those ones.

None of the 33+ bug reports I filed are related to your reports.

The most important (in terms of severity, gravity, criticality) bugs are those causing application hang and application crash:

Bug 247033: Generated content and display table- types hangs being one of them.


Then there are the bug reports which cause a lot of testcases to be failed:

Bug 235515: background-color applied to list-item generated box should not paint the marker box (list-style-position: outside) : causes 50 test failures

Bug 113728: Konqueror 4.6+ requires a minimum font-size of 2pt : causes 50 test failures

Bug 270546: line-height: [value] less than 3px or 2pt incorrectly rendered : causes 50 test failures

Bug 240026: background-* {image, position, repeat, attachment} properties should apply to table-column and table-colum-group elements : causes 25 test failures

Bug 197455: :first-letter pseudo-element should include punctuation character following first letter : causes 300 test failures; this one should be fixed in KDE 4.7/Konqueror 4.7

Bug 233899: width, min-width and max-width do not apply to table-row elements and to row groups elements (tbody, thead, tfoot) : causes 20 test failures



Overall, Konqueror/KHTML 4.6.5 does not do well in bi-directionality, in page media, in generated content. Other browsers also have more difficulties in those areas.

Test failures which could be considered as CSS 1 failures:

Bug 256759: [testcase] [CSS 2.1 Conformance] Background-repeat: repeat-x seems incorrectly rendered

Bug 256755: [testcase] [CSS 2.1 Conformance] Negative border-width in border shorthand declaration should be parsed as an error

c26-psudo-nest-000 first-letter testcase is definitely a CSS1 failure. In fact, Konqueror 4.6.5 also fails this sec26 first-letter test coming from CSS1 test suite

Bug 264909: font-family incorrectly selects the last of 2 generic font-family

There are a few other CSS1 test failures.

Gérard

GTalbot wrote:There are a few other CSS1 test failures.

Bug 189774: Malformed border shorthand declarations

Shorthand Properties (border) - maximum of 3 border subproperties : this one will be eventually included, inserted in the next CSS 2.1 test suite snapshot or RC

c23-first-line-000 : "first line" should be green

Gérard